Major Description
I'm running a query against JD's cluster and it's scanning much slower than expected - seems like a large amount of the data must be coming from MRS. The profile seems to agree with this, but there are no metrics showing how many rows/bytes are in MRS for each tablet.